With improved patient care, better diagnosis, and more treatment options after tumor recurrence, outcomes after fluorouracil (FU) -based treatment are expected to have improved over time in early-stage colon cancer. Data from 18,449 patients enrolled onto 21 phase III trials conducted from 1978 to 2002 were evaluated for potential differences in time to recurrence (TTR), time from recurrence to death (TRD), and overall survival (OS) with regard to FU-based adjuvant regimens.
Trials were predefined as old versus newer era using initial accrual before or after 1995. Outcomes were compared between patients enrolled onto old- or newer-era trials, stratified by stage.
Within the first 3 years, recurrence rates were lower in newer- versus old-era trials for patients with stage II disease, with no differences among those with stage III disease. Both TRD and OS were significantly longer in newer-era trials overall and within each stage. The lymph node (LN) ratio (ie, number of positive nodes divided by total nodes harvested) in those with stage III disease declined over time. TTR improved slightly, with larger number of LNs examined in both stages.
Improved TRD in newer trials supports the premise that more aggressive intervention (oxaliplatin- and irinotecan-based chemotherapy and/or surgery for recurrent disease) improves OS for patients previously treated in the adjuvant setting. Lower recurrence rates with identical treatments in those with stage II disease enrolled onto newer-era trials reflect stage migration over time, calling into question historical data related to the benefit of FU-based adjuvant therapy in such patients.
The association of deficient DNA mismatch repair (dMMR) with prognosis in patients with colon cancer treated with adjuvant fluorouracil, leucovorin, and oxaliplatin (FOLFOX) chemotherapy remains unknown.
Patients and Methods
Resected, stage III colon carcinomas from patients (N = 2,686) randomly assigned to FOLFOX ± cetuximab (North Central Cancer Treatment Group N0147 trial) were analyzed for mismatch repair (MMR) protein expression and mutations in BRAFV600E (exon 15) and KRAS (codons 12 and 13). Association of biomarkers with disease-free survival (DFS) was determined using Cox models. A validation cohort (Cancer and Leukemia Group B 88903 trial) was used.
dMMR was detected in 314 (12%) of 2,580 tumors, of which 49.3% and 10.6% had BRAFV600E or KRAS mutations, respectively. MMR status was not prognostic overall (adjusted hazard ratio [HR], 0.82; 95% CI, 0.64 to 1.07; P = .14), yet significant interactions were found between MMR and primary tumor site (Pinteraction = .009) and lymph node category (N1 v N2; Pinteraction = .014). Favorable DFS was observed for dMMR versus proficient MMR proximal tumors (HR, 0.71; 95% CI, 0.53 to 0.94; P = .018) but not dMMR distal tumors (HR, 1.71; 95% CI, 0.99 to 2.95; P = .056), adjusting for mutations and covariates. Any survival benefit of dMMR was lost in N2 tumors. Mutations in BRAFV600E (HR, 1.37; 95% CI, 1.08 to 1.70; P = .009) or KRAS (HR, 1.44; 95% CI, 1.21 to 1.70; P < .001) were independently associated with worse DFS. The observed MMR by tumor site interaction was validated in an independent cohort of stage III colon cancers (Pinteraction = .037).
The prognostic impact of MMR depended on tumor site, and this interaction was validated in an independent cohort. Among dMMR cancers, proximal tumors had favorable outcome, whereas distal or N2 tumors had poor outcome. BRAF or KRAS mutations were independently associated with adverse outcome.
Our study compares the outcomes of men and women with early stage colon cancer by analyzing the ACCENT database. Overall, men experienced inferior prognoses when compared with women for time to recurrence after adjusting for age, stage, and treatment. Sex was not a predictive factor of treatment efficacy. In exploratory analyses, worse outcomes in men were more prominent in older patients, but the stage of disease and type of adjuvant regimen did not modify the prognostic value of sex.
To compare long-term outcomes between men and women in a large cohort of clinical trial participants with early-stage colon cancer, specifically by examining whether the prognostic effect of sex varies based on age, stage of disease, and type of adjuvant therapy received.
A pooled analysis of individual patient data from 33,345 patients with colon cancer enrolled in 24 phase III studies of various adjuvant systemic therapies was conducted. Chemotherapy consisted of (1) fluorouracil (5-FU), (2) 5-FU variations, (3) 5-FU plus oxaliplatin, (4) 5-FU plus irinotecan, or (5) oral fluoropyrimidine-based regimens. The primary endpoint was disease-free survival; secondary endpoints included overall survival and time to recurrence. Stratified Cox models were used to assess the effect of sex on outcomes. Multivariate models were used to assess adjusted effects and to explore the interaction among sex and other factors.
A total of 18,244 (55%) men and 15,101 (45%) women were included. In the entire cohort, the median age was 61 years; 91% (24,868) were white; 31% (10,347) and 69% (22,964) had stage I/II and III disease, respectively. Overall, men had inferior prognoses when compared with women for time to recurrence (hazard ratio [HR] 1.05 [95% CI, 1.01–1.09]) and other endpoints after adjusting for age, stage, and treatment. Sex was not a predictive factor of treatment efficacy (P for interaction between sex and treatment when adjusting for age and stage were .40, .67, and .77 for disease-free survival, overall survival, and time to recurrence, respectively). In exploratory analyses, worse outcomes in men were more prominent in the older patients when adjusting for stage and treatment (HR 1.08 in age ≤ 65 years vs. HR 1.18 in age > 65 years; interaction P = .016 for disease-free survival). The stage of disease and type of adjuvant regimen did not modify the prognostic value of sex.
Sex is a modest independent prognostic marker for patients with early-stage colon cancer, particularly in older patients.
Sex; Colon cancer; Survival; Outcomes
Prior studies have suggested that patients with stage II/III colon cancer receive similar benefit from intravenous (IV) fluoropyrimidine adjuvant therapy regardless of age. Combination regimens and oral fluorouracil (FU) therapy are now standard. We examined the impact of age on colon cancer recurrence and mortality after adjuvant therapy with these newer options.
Patients and Methods
We analyzed 11,953 patients age < 70 and 2,575 age ≥ 70 years from seven adjuvant therapy trials comparing IV FU with oral fluoropyrimidines (capecitabine, uracil, or tegafur) or combinations of fluoropyrimidines with oxaliplatin or irinotecan in stage II/III colon cancer. End points were disease-free survival (DFS), overall survival (OS), and time to recurrence (TTR).
In three studies comparing oxaliplatin-based chemotherapy with IV FU, statistically significant interactions were not observed between treatment arm and age (P interaction = .09 for DFS, .05 for OS, and .36 for TTR), although the stratified point estimates suggested limited benefit from the addition of oxaliplatin in elderly patients (DFS hazard ratio [HR], 0.94; 95% CI, 0.78 to 1.13; OS HR, 1.04; 95% CI, 0.85 to 1.27). No significant interactions by age were detected with oral fluoropyrimidine therapy compared with IV FU; noninferiority was supported in both age populations.
Patients age ≥ 70 years seemed to experience reduced benefit from adding oxaliplatin to fluoropyrimidines in the adjuvant setting, although statistically, there was not a significant effect modification by age, whereas oral fluoropyrimidines retained their efficacy.
Background. A phase II design with an option for direct assignment (stop randomization and assign all patients to experimental treatment based on interim analysis, IA) for a predefined subgroup was previously proposed. Here, we illustrate the modularity of the direct assignment option by applying it to the setting of two predefined subgroups and testing for separate subgroup main effects. Methods. We power the 2-subgroup direct assignment option design with 1 IA (DAD-1) to test for separate subgroup main effects, with assessment of power to detect an interaction in a post-hoc test. Simulations assessed the statistical properties of this design compared to the 2-subgroup balanced randomized design with 1 IA, BRD-1. Different response rates for treatment/control in subgroup 1 (0.4/0.2) and in subgroup 2 (0.1/0.2, 0.4/0.2) were considered. Results. The 2-subgroup DAD-1 preserves power and type I error rate compared to the 2-subgroup BRD-1, while exhibiting reasonable power in a post-hoc test for interaction. Conclusion. The direct assignment option is a flexible design component that can be incorporated into broader design frameworks, while maintaining desirable statistical properties, clinical appeal, and logistical simplicity.
By using data from North Central Cancer Treatment Group Phase III Trial N0147, a randomized adjuvant trial of patients with stage III colon cancer, we assessed the relationship between smoking and cancer outcomes, disease-free survival (DFS), and time to recurrence (TTR), accounting for heterogeneity by patient and tumor characteristics.
Patients and Methods
Before random assignment to infusional fluorouracil, leucovorin, and oxaliplatin (FOLFOX) or FOLFOX plus cetuximab, 1,968 participants completed a questionnaire on smoking history and other risk factors. Cox models assessed the association between smoking history and the primary trial outcome of DFS (ie, time to recurrence or death), as well as TTR, adjusting for other clinical and patient factors. The median follow-up was 3.5 years among patients who did not experience events.
Compared with never-smokers, ever smokers experienced significantly shorter DFS (3-year DFS proportion: 70% v 74%; hazard ratio [HR], 1.21; 95% CI, 1.02 to 1.42). This association persisted after multivariate adjustment (HR, 1.23; 95% CI, 1.02 to 1.49). There was significant interaction in this association by BRAF mutation status (P = .03): smoking was associated with shorter DFS in patients with BRAF wild-type (HR, 1.36; 95% CI, 1.11 to 1.66) but not BRAF mutated (HR, 0.80; 95% CI, 0.50 to 1.29) colon cancer. Smoking was more strongly associated with poorer DFS in those with KRAS mutated versus KRAS wild-type colon cancer (HR, 1.50 [95% CI, 1.12 to 2.00] v HR, 1.09 [95% CI, 0.85 to 1.39]), although interaction by KRAS mutation status was not statistically significant (P = .07). Associations were comparable in analyses of TTR.
Overall, smoking was significantly associated with shorter DFS and TTR in patients with colon cancer. These adverse relationships were most evident in patients with BRAF wild-type or KRAS mutated colon cancer.
Frequently a biomarker capable of defining a patient population with enhanced response to an experimental agent is not fully validated with a known threshold at the start of a phase II trial. When such candidate predictive markers are evaluated and/or validated retrospectively, over-accrual of patients less likely to benefit from the regimen may result, leading to underpowered analyses or sub-optimal patient care.
We propose an adaptive randomized phase II study design incorporating prospective biomarker threshold identification (or non-identification), possible early futility stopping, potential mid-trial accrual restriction to marker-positive subjects, and final marker and treatment evaluation in the patient population identified as most likely to benefit.
An interim analysis is used to determine whether an initially unselected trial should stop early for futility, continue without a promising marker, or adapt accrual and resize (up to a pre-determined maximum) according to a promising biomarker. Final efficacy analyses are performed in the target population identified at the interim as most likely to benefit from the experimental regimen. Simulation studies demonstrate control of false-positive error rates, power, reduced average sample size, and other favorable aspects.
The design performs well at identifying a truly predictive biomarker at interim analysis, and subsequently restricting accrual to patients most likely to benefit from the experimental treatment. Type I and type II error rates are adequately controlled by restricting the range of marker prevalence via the candidate thresholds, and by careful consideration of the timing of interim analysis.
In situations where identification and validation of a naturally continuous biomarker are desired within a randomized phase II trial, the design presented herein offers a potential solution.
Adaptive design; randomized clinical trial; threshold identification; predictive biomarker; phase II trial; interim futility analysis
In the current era of stratified medicine and biomarker-driven therapies, the focus has shifted from predictions based on the traditional anatomic staging systems to guide the choice of treatment for an individual patient to an integrated approach using the genetic makeup of the tumor and the genotype of the patient. The clinical trial designs utilized in the developmental pathway for biomarkers and biomarker-directed therapies from discovery to clinical practice are rapidly evolving. While several issues need careful consideration, two critical issues that surround the validation of biomarkers are the choice of the clinical trial design (which is based on the strength of the preliminary evidence and marker prevalence), and biomarker assay related issues surrounding the marker assessment methods such as the reliability and reproducibility of the assay. In this review, we focus on trial designs aiming at personalized medicine in the context of early phase trials for initial marker validation, as well as in the context of larger definitive trials. Designs for biomarker validation are broadly classified as retrospective (i.e., using data from previously well-conducted randomized controlled trials (RCTs) versus prospective (enrichment, all-comers, hybrid or adaptive). We believe that the systematic evaluation and implementation of these design strategies are essential to accelerate the clinical validation of biomarker guided therapy.
All-comers design; adaptive design; biomarker; enrichment design; hybrid design; randomized controlled trial (RCT)
Phase II clinical trials aim to identify promising experimental regimens for further testing in phase III trials. In this review article, we focus on phase II designs for initial predictive biomarker validation to determine if a drug should be developed for an unselected patient population or for a biomarker-defined patient subset only. Several prospective designs for biomarker-directed therapy have been proposed, differing primarily in the study population, or randomization scheme, or both. The design choice is driven by scientific rationale, marker prevalence, strength of preliminary evidence, assay performance, and turn-around times for marker assessment. The enrichment design is most appropriate when compelling preliminary evidence suggests treatment benefit in only certain marker-defined subgroups, the all-comers design is useful when preliminary evidence regarding treatment effects in marker subgroups is unclear, and adaptive designs have the most potential in the setting of multiple treatment options and multiple marker-defined subgroups. We recently proposed a 2-stage phase II design that has the option for direct assignment (i.e., stop randomization and assign all patients to the experimental arm in Stage 2) based on interim analysis (IA) results. This design recognizes the need for randomization but also acknowledges the possibility of promising but inconclusive results after pre-planned IA. Simulation studies demonstrated that the direct assignment-option design has minimal power loss, marginal increase in type I error rates, and reasonable robustness to population shift effects. Systematic evaluation and implementation of these design strategies in the phase II setting is essential for accelerating the clinical validation of biomarker guided-therapy.
All-comers design; adaptive design; biomarker; direct assignment; enrichment design; interim analysis
Quality-adjusted life year (QALY) estimation is a well-known but little used technique to compare survival adjusted for complications. Lack of calibration and interpretation guidance hinders implementation of QALY analyses.
We conducted simulation studies to assess the impact of differences in survival, toxicity rates, and utility values on QALY results.
Survival comparisons used both log-rank and Wilcoxon testing. We examined power considerations for a North Central Cancer Treatment Group Phase III lung cancer clinical trial (89-20-52).
Sample sizes of 100 events per treatment have low power to generate a statistically significant difference in QALYs unless the toxicity rate is 44% higher in one arm. For sample sizes of 200 per arm and equal survival times, toxicity needs to be at least 38% more in one arm for the result to be statistically significant, using a utility of 0.3 for days with toxicity. Sample sizes of 300 (500)/arm provide 80% power if there is a 31% (25%) toxicity difference. If the overall survival hazard ratio between the two treatment arms is 1.25, then samples of at least 150 patients and 13% increased toxicity are necessary to have 80% power to detect QALY differences. In study 89-20-52, there was only 56% power to determine the statistical significance of the observed QALY differences, clarifying the enigmatic conclusion of no statistically significant difference in QALY despite an observed 14.5% increase in toxicity between treatments.
This calibration allows researchers to interpret the clinical significance of QALY analyses and facilitates QALY inclusion in clinical trials through improved study design.
QALY; quality-adjusted life year; Q-TWiST; QOL; quality of life; simulation
In this commentary we discuss several challenges that are of current relevance to the design of clinical trials in oncology. We argue that the compartmentalization of trials into the three standard phases, with non overlapping aims, is not necessary and in fact may slow the clinical development of agents. Combined phase I/II trials and/or phase I trials that at minimum collect efficacy data and more optimally include a preliminary measure of efficacy in dosing determination should be more widely utilized. Similarly, we posit that randomized phase II trials should be used more frequently, as opposed to the traditional historical single arm phase II trial that usually does not have a valid comparison group. The use of non binary endpoints is a simple modification that can improve the efficiency of early phase trials. The heterogeneity in scientific goals and contexts in early phase oncology trials is considerable, and the potential to improve the design to match these goals is great. Our overall premise is that the potential benefits associated with the oncology clinical trial community moving away from the one size fits all paradigm of trial design are great, and that more flexible and efficient designs tailored to match the goals of each study are currently available and being used successfully.
Efficient designs; flexible adaptive designs; summarization of information; randomized phase II; phase I/II
Multiple GWAS have identified several susceptibility variants for colon cancer at 8q24. However, the functional roles of these variants have yet to be elucidated. Here, we evaluated the potential role of these markers in tumor progression and examined association with commonly observed structural abnormalities in this region, c-MYC amplification and chromosome fragility at FRA8C and FRA8D. We first replicated the previously reported association by testing 1178 cases and 1009 clinic-based controls with eight markers localized to three specific regions at 8q24. We observed significant associations with colon cancer risk with markers rs13254738 (ordinal OR=0.82, 95% CI=0.072-0.94, Ptrend=0.0037) and rs6983267 (ordinal OR=1.17, 95% CI=1.03-1.32, Ptrend=0.013). Survival analysis was performed using a separate set of 460 cases to evaluate the clinical significance of these markers. Overall, univariate analysis did not detect survival differences for any of the markers. We also tested a subset of the 460 cases (N=380) for structural abnormalities at or near the c-MYC locus using FISH analysis. Furthermore, we evaluated a small number of cases homozygous for the rs6983267 alleles to test for differences in fragile site induction. None of the 8q markers correlated with amplification at the c-MYC locus as detected by FISH, and no clear pattern of breakage was observed at the FRA8C and FRA8D sites. In this study, we confirm the association for several SNPs at 8q24 in colon cancer but have not detected any structural role relating to c-MYC amplification or chromosomal fragility. Finally, these risk alleles do not appear to be associated with survival.
8q; SNP; association; survival; FISH; fragile site; c-MYC
Although the importance of obesity in colon cancer risk and outcome is recognized, the association of body mass index (BMI) with DNA mismatch repair (MMR) status is unknown.
Patients and Methods
BMI (kg/m2) was determined in patients with TNM stage II or III colon carcinomas (n = 2,693) who participated in randomized trials of adjuvant chemotherapy. The association of BMI with MMR status and survival was analyzed by logistic regression and Cox models, respectively.
Overall, 427 (16%) tumors showed deficient MMR (dMMR), and 630 patients (23%) were obese (BMI ≥ 30 kg/m2). Obesity was significantly associated with younger age (P = .021), distal tumor site (P = .012), and a lower rate of dMMR tumors (10% v 17%; P < .001) compared with normal weight. Obesity remained associated with lower rates of dMMR (odds ratio, 0.57; 95% CI, 0.41 to 0.79; P < .001) after adjusting for tumor site, stage, sex, and age. Among obese patients, rates of dMMR were lower in men compared with women (8% v 13%; P = .041). Obesity was associated with higher recurrence rates (P = .0034) and independently predicted worse disease-free survival (DFS; hazard ratio [HR], 1.37; 95% CI, 1.14 to 1.64; P = .0010) and overall survival (OS), whereas dMMR predicted better DFS (HR, 0.59; 95% CI, 0.47 to 0.74; P < .001) and OS. The favorable prognosis of dMMR was maintained in obese patients.
Colon cancers from obese patients are less likely to show dMMR, suggesting obesity-related differences in the pathogenesis of colon cancer. Although obesity was independently associated with adverse outcome, the favorable prognostic impact of dMMR was maintained among obese patients.
It is widely accepted that traditional models of clinical investigation are becoming unsustainable in oncology and that trials must become more efficient in matching effective treatments to the patients most likely to benefit. In 2008, gastrointestinal oncologists from many countries began a collaboration to improve the design and conduct of clinical trials in their field, through the auspices of a French/U.S. charitable foundation, ARCAD. Whether this model of academic collaboration will be judged a success will depend on the quality of its scientific output during the next few years and whether this output, alongside that of other scientists, groups, and institutions, ultimately leads to more efficient trials and improved treatment options for patients.
Gastrointestinal neoplasms; Clinical trials; Drugs; Investigational; Database; Data interpretation; Statistical
Traditionally, phase II trials have been conducted as single-arm trials to compare the response probabilities between an experimental therapy and a historical control. Historical control data, however, often have a small sample size, are collected from a different patient population, or use a different response assessment method, so that a direct comparison between a historical control and an experimental therapy may be severely biased. Randomized phase II trials entering patients prospectively to both experimental and control arms have been proposed to avoid any bias in such cases. The small sample sizes for typical phase II clinical trials imply that the use of exact statistical methods for their design and analysis is appropriate. In this paper, we propose two-stage randomized phase II trials based on Fisher’s exact test which does not require specification of the response probability of the control arm for testing. Through numerical studies, we observe that the proposed method controls the type I error accurately and maintains a high power. If we specify the response probabilities of the two arms under the alternative hypothesis, we can identify good randomized phase II trial designs by adopting the Simon’s minimax and optimal design concepts that were developed for single-arm phase II trials.
Fisher’s exact test; Minimax design; Optimal design; Two-stage design; Unbalanced allocation
Symptoms and complications of metastatic colorectal cancer (mCRC) differ by metastatic sites. There is a paucity of prospective survival data for patients with peritoneal carcinomatosis colorectal cancer (pcCRC). We characterized outcomes of patients with pcCRC enrolled onto two prospective randomized trials of chemotherapy and contrasted that with other manifestations of mCRC (non-pcCRC).
A total of 2,095 patients enrolled onto two prospective randomized trials were evaluated for overall survival (OS) and progression-free survival (PFS). A Cox proportional hazard model was used to assess the adjusted associations.
The characteristics of the pcCRC group (n = 364) were similar to those of the non-pcCRC patients in median age (63 v 61 years, P = .23), sex (57% males v 61%, P = .23), and performance status (Eastern Cooperative Oncology Group performance status 0 or 1 94% v 96%, P = .06), but differed in frequency of liver (63% v 82%, P < .001) and lung metastases (27% v 34%, P = .01). Median OS (12.7 v 17.6 months, hazard ratio [HR] = 1.3; 95% CI, 1.2 to 1.5; P < .001) and PFS (5.8 v 7.2 months, HR = 1.2; 95% CI, 1.1 to 1.3; P = .001) were shorter for pcCRC versus non-pcCRC. The unfavorable prognostic influence of pcCRC remained after adjusting for age, PS, liver metastases, and other factors (OS: HR = 1.3, P < .001; PFS: HR = 1.1, P = .02). Infusional fluorouracil, leucovorin, and oxaliplatin was superior to irinotecan, leucovorin, and fluorouracil as a first-line treatment among pcCRC (HR for OS = 0.62, P = .005) and non-pcCRC patients (HR = 0.66, P < .001).
pcCRC is associated with a significantly shorter OS and PFS as compared with other manifestations of mCRC. Future trials for mCRC should consider stratifying on the basis of pcCRC status.
Stool DNA testing is a new approach to colorectal cancer detection. Few data are available from the screening setting.
To compare stool DNA and fecal blood testing for detection of screen-relevant neoplasia (curable-stage cancer, high-grade dysplasia, or adenomas >1 cm).
Blinded, multicenter, cross-sectional study.
Communities surrounding 22 participating academic and regional health care systems in the United States.
4482 average-risk adults.
Fecal blood and DNA markers. Participants collected 3 stools, smeared fecal blood test cards and used same-day shipment to a central facility. Fecal blood cards (Hemoccult and HemoccultSensa, Beckman Coulter, Fullerton, California) were tested on 3 stools and DNA assays on 1 stool per patient. Stool DNA test 1 (SDT-1) was a precommercial 23-marker assay, and a novel test (SDT-2) targeted 3 broadly informative markers. The criterion standard was colonoscopy.
Sensitivity for screen-relevant neoplasms was 20% by SDT-1, 11% by Hemoccult (P = 0.020), 21% by HemoccultSensa (P = 0.80); sensitivity for cancer plus high-grade dysplasia did not differ among tests. Specificity was 96% by SDT-1, compared with 98% by Hemoccult (P < 0.001) and 97% by HemoccultSensa (P = 0.20). Stool DNA test 2 detected 46% of screen-relevant neoplasms, compared with 16% by Hemoccult (P < 0.001) and 24% by HemoccultSensa (P < 0.001). Stool DNA test 2 detected 46% of adenomas 1 cm or larger, compared with 10% by Hemoccult (P < 0.001) and 17% by HemoccultSensa (P < 0.001). Among colonoscopically normal patients, the positivity rate was 16% with SDT-2, compared with 4% with Hemoccult (P = 0.010) and 5% with HemoccultSensa (P = 0.030).
Stool DNA test 2 was not performed on all subsets of patients without screen-relevant neoplasms. Stools were collected without preservative, which reduced detection of some DNA markers.
Stool DNA test 1 provides no improvement over HemoccultSensa for detection of screen-relevant neoplasms. Stool DNA test 2 detects significantly more neoplasms than does Hemoccult or HemoccultSensa, but with more positive results in colonoscopically normal patients. Higher sensitivity of SDT-2 was particularly apparent for adenomas.
Assessing between-study variability in the context of conventional random-effects meta-analysis is notoriously difficult when incorporating data from only a small number of historical studies. In order to borrow strength, historical and current data are often assumed to be fully homogeneous, but this can have drastic consequences for power and Type I error if the historical information is biased. In this paper, we propose empirical and fully Bayesian modifications of the commensurate prior model (Hobbs et al., 2011) extending Pocock (1976), and evaluate their frequentist and Bayesian properties for incorporating patient-level historical data using general and generalized linear mixed regression models. Our proposed commensurate prior models lead to preposterior admissible estimators that facilitate alternative bias-variance trade-offs than those offered by pre-existing methodologies for incorporating historical data from a small number of historical studies. We also provide a sample analysis of a colon cancer trial comparing time-to-disease progression using a Weibull regression model.
clinical trials; historical controls; meta-analysis; Bayesian analysis; survival analysis; correlated data
Although obesity is an established risk factor for developing colon cancer, its prognostic impact and relationship to patient sex in colon cancer survivors remains unclear.
We examined the prognostic and predictive impact of the body mass index (BMI) in stage II and III colon carcinoma patients (N=25,291) within the Adjuvant Colon Cancer Endpoints (ACCENT) database. BMI was measured at enrollment in randomized trials of 5-fluorouracil-based adjuvant chemotherapy. Association of BMI with time-to-recurrence (TTR), disease-free (DFS), and overall survival (OS) was determined using Cox regression models. Statistical tests were two-sided.
During a median follow-up of 7.8 years, obese and underweight patients had significantly poorer survival compared to overweight or normal-weight patients. In a multivariable analysis, the adverse prognostic impact of BMI was seen in men but not in women (Pinteraction=.0129). Men with class 2/3 obesity (BMI ≥35.0 kg/m2) had a statistically significant reduction in DFS [HR= 1.16 (1.01, 1.33), p=.0297] vs normal weight patients Underweight patients had a significantly shorter TTR and reduced DFS [HR=1.18 (1.09, 1.28), p<.0001] that was stronger among men [HR= 1.31 (1.15, 1.50), p<.0001] than women [1.11 (1.01, 1.23), .0362; Pinteraction=.0340). BMI was not predictive of adjuvant treatment benefit.
Obesity and underweight status were independently associated with inferior outcomes in colon cancer patients treated in adjuvant chemotherapy trials.
Colorectal cancer (CRC) risk is partly conferred by common, low-penetrance single nucleotide polymorphisms (SNPs). We hypothesized that these SNPs are associated with outcomes in metastatic CRC.
Six candidate SNPs from 8q24, 10p14, 15q13, 18q21 were investigated for their association with response rate (RR), time to progression (TTP) and overall survival (OS) among 524 patients treated on a phase III clinical trial of first-line chemotherapy for metastatic CRC.
rs10795668 was weakly associated with TTP (p = 0.02), but not RR or OS. No other SNPs carried statistically significant HRs for any of the primary outcomes (RR, TTP or OS).
Common low-penetrance CRC risk SNPs were not associated with outcomes among patients with metastatic CRC.
Assessing impact of poor accrual on premature trial closure requires a relevant metric. We propose defining accrual sufficiency on apparent ability to address primary endpoints (PE) rather than attaining accrual targets.
All phase III trials open January 1, 1993, to December 31, 2002, by five U.S. oncology Clinical Trials Cooperative Groups (CTCG) were evaluated for accrual sufficiency and scientific results. Sufficient accrual included meeting accrual target, CTCGs documentation attesting adequate accrual, or conclusive results at interim analysis; insufficient accrual included poor accrual as cited closure reason or other reasons rendering a trial unable to address its primary endpoints. Closure rates based on our accrual sufficiency definition are compared with rates of meeting accrual targets and addressing the primary endpoints. A percentage of target accrual above which trials commonly answer the intended scientific question was identified to serve as an alternative to meeting full target accrual in designating accrual success.
Of 238 eligible trials, 158 (66%) closed with sufficient accrual. Among 80 trials with insufficient accrual, 70 (29%) closed specifically because of poor accrual. Inadequate accrual rates are overemphasized when defining accrual success solely by meeting accrual targets. Nearly 75% of trials conclusively addressed the primary endpoints with positive results in 39% of trials. Exceeding 80% of target accrual serves as a reliable proxy for answering the intended scientific question.
Approximately one third of phase III trials closed with insufficient accrual to address the primary endpoints, primarily due to poor accrual. Defining accrual sufficiency broader than meeting accrual targets represents a fairer account of trial closures.
A major challenge for randomized phase III oncology trials is the frequent low rates of patient enrollment, resulting in high rates of premature closure due to insufficient accrual.
We conducted a pilot study to determine the extent of trial closure due to poor accrual, feasibility of identifying trial factors associated with sufficient accrual, impact of redesign strategies on trial accrual, and accrual benchmarks designating high failure risk in the clinical trials cooperative group (CTCG) setting.
A subset of phase III trials opened by five CTCGs between August 1991 and March 2004 was evaluated. Design elements, experimental agents, redesign strategies, and pretrial accrual assessment supporting accrual predictions were abstracted from CTCG documents. Percent actual/predicted accrual rate averaged per month was calculated. Trials were categorized as having sufficient or insufficient accrual based on reason for trial termination. Analyses included univariate and bivariate summaries to identify potential trial factors associated with accrual sufficiency.
Among 40 trials from one CTCG, 21 (52.5%) trials closed due to insufficient accrual. In 82 trials from five CTCGs, therapeutic trials accrued sufficiently more often than nontherapeutic trials (59% vs 27%, p = 0.05). Trials including pretrial accrual assessment more often achieved sufficient accrual than those without (67% vs 47%, p = 0.08). Fewer exclusion criteria, shorter consent forms, other CTCG participation, and trial design simplicity were not associated with achieving sufficient accrual. Trials accruing at a rate much lower than predicted (<35% actual/predicted accrual rate) were consistently closed due to insufficient accrual.
This trial subset under-represents certain experimental modalities. Data sources do not allow accounting for all factors potentially related to accrual success.
Trial closure due to insufficient accrual is common. Certain trial design factors appear associated with attaining sufficient accrual. Defining accrual benchmarks for early trial termination or redesign is feasible, but better accrual prediction methods are critically needed. Future studies should focus on identifying trial factors that allow more accurate accrual predictions and strategies that can salvage open trials experiencing slow accrual.
Postoperative outcomes of patients undergoing laparoscopic-assisted colectomy (LAC) have shown modest improvements in recovery but only minimal differences in quality of life (QOL) compared with open colectomy. We therefore sought to assess the effect of LAC on QOL in the short and long term, using individual item analysis of multi-item QOL assessments.
QOL variables were analyzed in 449 randomized patients from the COST trial 93-46-53 (INT 0146). Both cross-sectional single-time and change from baseline assessments were run at day 2, week 2, month 2, and month 18 postoperatively in an intention-to-treat analysis using Wilcoxon rank-sum tests. Stepwise regression models were used to determine predictors of QOL.
Of 449 colon cancer patients, 230 underwent LAC and 219 underwent open colectomy. Subdomain analysis revealed a clinically moderate improvement from baseline for LAC in total QOL index at 18 months (P = 0.02) as well as other small symptomatic improvements. Poor preoperative QOL as indicated by a rating scale of ≤ 50 was an independent predictor of poor QOL at 2 months postoperatively. QOL variables related to survival were baseline support (P = 0.001) and baseline outlook (P = 0.01).
Eighteen months after surgery, any differences in quality of life between patients randomized to LAC or open colectomy favored LAC. However, the magnitude of the benefits was small. Patients with poor preoperative QOL appear to be at higher risk for difficult postoperative courses, and may be candidates for enhanced ancillary services to address their particular needs.
The prognostic value of actual tumor measurements versus World Health Organization criteria as three-level variables (incorporating stable disease) and two-level variables (stable disease not included in “response”) at 12 and 24 weeks as predictors of survival in a phase III trial in metastatic colorectal cancer was explored. In this trial, the use of actual tumor measurement, or following tumor status beyond 12 weeks, did not improve survival prediction compared with a single three-level response assessment at 12 weeks, suggesting that 12-week tumor status could be an appropriate phase II trial endpoint in metastatic colorectal cancer.
After completing this course, the reader will be able to:
Explain the difference in survival prediction between response criteria (WHO) when used as a two-level variable (CR/PR vs. other) and as a three-level variable (CR/PR vs. SD vs. PD).Describe the limited benefit of using actual tumor measurements over traditional criteria (as a three-level variable) in predicting survival in colorectal cancer.
This article is available for continuing medical education credit at CME.TheOncologist.com
We explored the prognostic value of actual tumor measurements (TM) versus World Health Organization (WHO) criteria as three-level (responder, stable, and progression) and two-level (responder and non-responder) variables at 12 and 24 weeks as predictors of survival in Intergroup Trial N9741, a phase III trial in metastatic colorectal cancer (CRC).
All patients with measurable disease (N = 1,188) were included. The percentage changes in TM from baseline to 12 and 24 weeks were calculated. The prognostic values of TM versus WHO criteria (as three- and two-level variables) at 12 and 24 weeks were compared, using Cox models for overall survival (OS) in a landmark analysis, adjusting for baseline tumor size, performance status, and treatment arm.
Tumor status at 12 weeks by WHO criteria (three or two levels) or actual TM were all strongly associated with OS. Actual TM provided no meaningful additional benefit compared with the three-level WHO criteria. Tumor status at 24 weeks was also strongly associated with survival, but added no additional prognostic value compared with the 12-week assessment. At 12 weeks, actual TM improved prognostic characterization of patients with WHO status of response, but provided no additional value in patients with stable disease or progression.
In N9741, the use of actual TM, or following tumor status beyond 12 weeks, did not improve survival prediction compared with a single three-level response assessment at 12 weeks, suggesting that 12-week tumor status could be an appropriate phase II trial endpoint in metastatic CRC.
Colorectal cancer; Tumor status; Endpoints; Clinical trial
With the advent of targeted therapies, biomarkers provide a promising means of individualizing therapy through an integrated approach to prediction using the genetic makeup of the disease and the genotype of the patient. Biomarker validation has therefore become a central topic of discussion in the field of medicine, primarily due to the changing landscape of therapies for treatment of a disease and these therapies purported mechanism(s) of action.
In this report, we discuss the merits and limitations of some of the clinical trial designs for predictive biomarker validation using examples from ongoing or completed clinical trials.
The designs are broadly classified as retrospective (i.e., using data from previously well-conducted randomized controlled trials (RCT)) versus prospective (enrichment or targeted, unselected or all-comers, hybrid, and adaptive analysis). We discuss some of these designs in the context of real trials.
Well-designed retrospective analysis of prospective RCT can bring forward effective treatments to marker defined subgroup of patients in a timely manner. An example is the KRAS gene status in colorectal cancer – the benefit from cetuximab and panitumumab was demonstrated to be restricted to patients with wild type status based on prospectively specified analyses using data from previously conducted RCTs. Prospective enrichment designs are appropriate when compelling preliminary evidence suggests that not all patients will benefit from the study treatment under consideration; however, this may sometimes leave questions unanswered. An example is the established benefit of trastuzumab as adjuvant therapy for breast cancer; a clear definition of HER2-positivity and the assay reproducibility have, however, remained unanswered. An all-comers design is optimal where preliminary evidence regarding treatment benefit and assay reproducibility is uncertain (e.g., EGFR expression and tyrosine kinase inhibitors in lung cancer), or to identify the most effective therapy from a panel of regimens (e.g., chemotherapy options in breast cancer).
The designs discussed here rest on the assumption that the technical feasibility, assay performance metrics, and the logistics of specimen collection are well established and that initial results demonstrate promise with regard to the predictive ability of the marker(s).
The choice of a clinical trial design is driven by a combination of scientific, clinical, statistical, and ethical considerations. There is no one size fits all solution to predictive biomarker validation.